The Clinical Study Of Intensive Insulin Therapy On Severe Sepsis

ObjectiveTo investigate the effect of intensive insulin therapy on severesepsis.MethodsIn a prospective, randomized, controlled study of adult patientsadmitted to our ICU, we studied patients who were considered to needintensive care for at least three days. On admission, one hundred andforty-two severe sepsis patients with hyperglycemia were randomlydividided into intensive insulin therapy group (maintenance of bloodglucose at a level between 4.4 and 6.1 mmol per liter) or conventionaltherapy group (insulin administered when the blood glucose level exceeded12 mmol per liter, with the infusion tapered when the level fell below10mmol per Liter and maintenance of glucose at a level between 10. 0and11.1 mmol per liter). Mortality in the ICU, mortality in the hospital,days weaning from mechanical ventilation, days in the ICU and in thehospital, new kidney injury and new hepatic injury during intensive carewere observed as outcome measures.ResultsThe days in the ICU(12±3) and in the hospital (19±6) and the daysweaning from mechanical ventilation in intensive insulin therapy groupwere significantly reduced than in the conventional therapy group (P＜0.001). The mortality in the ICU (22.5% in the intensive insulin therapygroup vs 42.3% in the conventional therapy group, P＜0.05) and in thehospital(36.6% vs 54.9% P＜0.05)were significantly reduced. Morever,morbidity was significantly reduced by the prevention of newly acquiredkidney injury(8.5% vs 19.7% P＜0.05).ConclusionInensive insulin therapy in severe sepsis patients who were withhyperglycemia and stayed in the ICU for at least three days reducesmortality in the ICU and hospital. In addition ,Inensive insulin therapyin severe sepsis patients was associated with reduced days in the ICU andhospital, and reduced days weaning from ventilation as well as themorbidity. Glucose levels in severesepsis patients should be controlledthrough implementation of intensive insulin therapy with the goal toachieve normoglycemia, regardless of a history of diabetes.Frequent monitoring is imperative to avoid hypoglycemia. Largemulticenter trials are needed to confirm these preliminary results.